Goodies for checkmk
- It is only tested under Debian Linux (12) 'bookworm' with check_mk 2.2.
- It comes with no warranty at all!!!
- Support will be not available.
- Contributions are welcome
- Download the script(s) you want to use (or clone the github-repo)
- Put it/them into the folder
/usr/lib/check_mk_agent/local/ - Make the script(s) executable (0755)
- Rescan the host in check_mk for new services
(Partially based on the gluster checkmk plugin ( https://exchange.checkmk.com/p/gluster ) from lpwevers.)
Requirements:
- glusterfs (of course)
- pidof
Description
It monitors glusterfs demons (running-state) and peers (connection-state) and volumes (needed healings). Also, the needed healings of a volume are available as a metric for this volume.
At the head of the script, you can configure some time periods and other values.
Note
The changes to the checkmk plugin from lpwevers are motivated as follows: The plugin immediately sets the service level to 'Critical', as soon as at least one file needs to be healed. However, this seems to be a normal and regular process of glusterfs - at least if you use it under proxmox, as I do - which in my observation should not immediately trigger an alert in the monitoring.
The 'gluster' script therefore takes over all checks (except for the quotas) from lpwevers´ plugin and supplements the volume healing check with the option of first triggering a warning and then a critical alert after a certain (and configurable) time or a certain (and configurable) count of healings.
To use this script, you have to disable lpwevers´ plugin.